• The Race to Save the Planet: Could machine learning bring algae into the sunlight?

The Race to Save the Planet: Could machine learning bring algae into the sunlight?

12 November 2021

Over the past weeks, the limiting climate change has become the main point of discussion as world leaders met for the 26th edition of the Conference of the Parties (COP26) in Glasgow. Although climate change has been a continuing phenomenon for centuries, the calls for immediate and urgent action have become louder and critical mitigation strategies are required to manage irreversible damage to our planet.

In their discussions, world leaders from over 100 countries made a vow to stop deforestation by the year 2030. While this is a noble cause, there are other strategies that could be effective in tackling the climate crisis.

This is where the argument for the use of algae, in conjunction with reforestation efforts, could be made. With advances in machine learning, through artificial neural networks and regression analysis, companies could harness emerging technology to nurture and encourage the production of newly formed algae blooms whilst limiting dangerous levels of algae production.

How algae can help manage the climate crisis

Algae are a diverse group of aquatic organisms that have the ability to conduct photosynthesis, thereby removing carbon dioxide from the atmosphere and converting it into oxygen. Although there are risks due to harmful algae blooms (species of phytoplankton that can have negatives impacts on marine life), the benefits of algae cultivation for the removal of carbon dioxide from the atmosphere are undeniable. It also has potential to be farmed for biofuels and fertiliser.  

AI-powered experiments with algae have seen controlled algae blooms remove 400 times more carbon dioxide than a tree from the atmosphere. If emerging Machine Learning technology can be harvested to predict adaptation and evolution of safe algae blooms species to rising temperatures, Global governments could focus on the slowing the speed of global warming to a level where the growth of the healthy blooms could be encouraged. This would then absorb more carbon dioxide than previously possible and, in turn, reduce the global temperature.

Predictive modelling using phytoplankton discovery

By making use of machine learning methods and technologies such as neural network models, we believe companies could focus their efforts on building an algae-based ecological predictive model with a mixture of both quantitative and qualitative data in programming languages which include Python and R. The ability to train the algorithms on a substantial amount of time-series input data, including hydrological data and images collected from polar-orbiting satellites, is of utmost importance if the necessary feature subsets are to be generated. A combination of the most relevant parameters, such as, cell concentration and water temperature, and an efficient selection process to identify the most important features from the data is required. Several models may need to be used to ensure the that the rate of adaptation is such that the healthy species, is outpacing harmful blooms, whilst the carbon dioxide absorption characteristics of the algae cells continues to at least remain at current levels.


Attempts to limit deforestation will be an ongoing struggle but the leveraging algae blooms could also make a significant impact on rising CO2 levels. We believe that machine learning models can help businesses exploit algae for the greater good of our planet.

While ongoing R&D developments within the environmental and ecological space are essential for climate change mitigation, the inevitable costs and risks associated with these developments cannot be ignored. Fortunately, machine learning projects that aim to solve a technological uncertainty fall within the criteria for R&D tax purposes and ultimately qualify for tax credits.

For more advice and guidance on the tax treatment of the novel technology you are engineering, please contact Eyad Hamouieh or Ashley Rawson.

Enter security code:
 Security code